Phycologia (2008) Volume 47 (4), 371–391 Published 3 July 2008

Pseudostriatella (Bacillariophyta): a description of a new araphid diatom genus based on observations of frustule and auxospore structure and 18S rDNA phylogeny

1 2 3 1 SHINYA SATO *, DAVID G. MANN ,SATOKO MATSUMOTO AND LINDA K. MEDLIN 1Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, D-27570 Bremerhaven, Germany 2Royal Botanic Garden, Edinburgh EH3 5LR, Scotland, United Kingdom 3Choshi Fisheries High School, 1-1-12 Nagatsuka Cho, Choshi City, Chiba, Japan

S. SATO, D.G. MANN,S.MATSUMOTO AND L.K. MEDLIN. 2008. Pseudostriatella (Bacillariophyta): a description of a new araphid diatom genus based on observations of frustule and auxospore structure and 18S rDNA phylogeny. Phycologia 47: 371–391. DOI: 10.2216/08-02.1

Pseudostriatella oceanica gen et. sp. nov. is a marine benthic diatom that resembles Striatella unipunctata in gross morphology, attachment to the substratum by a mucilaginous stalk and possession of septate girdle bands. In light microscopy, P. oceanica can be distinguished from S. unipunctata by plastid shape, absence of truncation of the corners of the frustule, indiscernible striation and absence of polar rimoportulae. With scanning electron microscopy, P. oceanica can be distinguished by a prominent but unthickened longitudinal hyaline area, pegged areolae, multiple marginal rimoportulae and perforated septum. The hyaline area differs from the sterna of most pennate diatoms in being porous toward its expanded ends; in this respect, it resembles the elongate annuli of some centric diatoms, such as Attheya and Odontella. 18S rDNA phylogeny places P. oceanica among the pennate diatoms and supports a close relationship between P. oceanica and S. unipunctata, but the genetic distance between them, coupled with the morphological differences, justifies separation at genus level. However, the affinity of the P. oceanica – S. unipunctata clade remains unresolved both in molecular and in morphological study. Both genera are only distantly related to Hyalosira and Grammatophora, despite similarities in frustule structure and growth habit, arguing against their inclusion in the same family. The auxospore is covered with series of transverse and longitudinal bands, but the structure and arrangement of these bands appear to be more similar to the properizonia of some centric diatoms than to the classic type of perizonium seen in other pennate diatoms; a few scales are also present. The differences between properizonia and perizonia are discussed.

KEY WORDS: 18S rDNA, Araphid diatom, Auxospore, Evolution, Fine structure, Morphology, Perizonium, Phylogeny, Pseudostriatella oceanica, Striatella, Taxonomy

INTRODUCTION high abundance, the defining features of the main groups of araphid diatoms are not fully established. Benthic diatoms are ubiquitous in shallow coastal environ- To obtain a more complete picture of the natural history ments and are one of the most taxonomically diverse of araphid diatoms, we have been collecting samples groups of organisms in estuarine ecosystems (Sullivan & worldwide from coastal regions. Recently we encountered Currin 2000). Because of their high primary production a new diatom that superficially resembled Striatella rates, benthic diatoms play an important role in the unipunctata (Lyngbye) Agardh. Scanning electron micros- functioning of benthic trophic webs in intertidal mudflats copy (SEM) revealed, however, that this diatom differed and shallow-water ecosystems of temperate to tropical from S. unipunctata in several features that are generally regions (Cahoon 1999; Underwood & Kromkamp 1999). used as taxonomic characters among araphid diatoms, Araphid pennate diatoms (diatoms with a sternum but including characteristics of the sternum, striae, areolae, lacking a raphe system; see Terminology) are important apical pore field, rimoportula and septum. Given these components of these coastal assemblages, particularly observations, together with information on the plastids and among communities attached to macrophytes and macro- 18S rDNA sequences, we conclude that the diatom should algae, animals, rocks and sand grains (Round et al. 1990). be described as a new genus, Pseudostriatella. Taxonomically, araphid diatoms have long been neglected, We have also been able to make detailed observations on perhaps because of their morphological simplicity; accord- the fine structure of auxospores produced spontaneously in ing to Round et al. (1990), ‘in many ways the classification monoclonal cultures. With the advent of electron micros- of the araphid group is the most difficult because unlike the copy, particularly SEM, information about auxospore centric series their valve structure is rather simple, and structure has greatly increased (e.g. Crawford 1974; Mann unlike the raphid series, the plastids and their arrangements 1982b; von Stosch 1982; Cohn et al. 1989; Kaczmarska et have few distinguishing features’. Thus, in spite of their al. 2000, 2001; Kobayashi et al. 2001; Schmid & Crawford 2001; Nagumo 2003; Sato et al. 2004, 2008a, b; Amato et al. * Corresponding author ([email protected]). 2005; Tiffany 2005; Toyoda et al. 2005, 2006; Trobajo et al.

371 372 Phycologia, Vol. 47 (4), 2008

2006; Poulı´cˇkova´ & Mann 2006; Poulı´cˇkova´ et al. 2007). was left at room temperature for c. 30 min; and (5) steps 1 However, although it has become clear that some aspects of and 2 were repeated several times to remove decomposition the fine structure of auxospores have phylogenetic signif- products. Cleaned frustules were then mounted in Mount- icance (e.g. Medlin & Kazcmarska 2004), there is still media (refractive index n20/D 5 1.50; Wako). insufficient information to reveal how the structure and For SEM examination, cleaned material was air-dried development of auxospores have evolved in the major onto coverslips. To observe auxospores, coverslips to which diatom groups, especially among the lineages of araphid the auxospore mother cells had already become attached pennate diatoms. Indeed, the only detailed information were immersed in 10% glutaraldehyde for 1 h at room available concerning araphid pennates is the account of temperature, then washed with distilled water, air-dried and Rhabdonema Ku¨tzing by von Stosch (1962, 1982) and the fixed to SEM stubs with carbon tape. For observations of SEM studies of Gephyria media Arnott (Sato et al. 2004), cells still attached to the substratum by mucilaginous stalks, Grammatophora marina (Lyngbye) Ku¨tzing (Sato et al. host plants were fixed with 10% glutaraldehyde for 2 h at 2008a) and Tabularia parva (Ku¨tzing) Williams & Round 4uC, rinsed with distilled water several times to remove the (Sato et al. 2008b). In the present study, we compare the glutaraldehyde, dehydrated using increasing concentrations auxospore fine structure in these diatoms with that of of t-butyl alcohol and freeze-dried using a JFD-310 Pseudostriatella oceanica and discuss the evolutionary instrument (JEOL). Freeze-dried specimens were attached relationships of Pseudostriatella. to the stub directly with carbon tape. All SEM specimens were coated with gold using an SC 500 sputter coater (Emscope). A QUANTA 200F (FEI) was used for SEM MATERIAL AND METHODS observation at an accelerating voltage of 3–10 kV and c. 10 mm working distance. All the images included in this paper are from cultured strains, except for those from Collections and cultures freeze-dried material (Figs 9–13). Captured images were Both natural specimens and clonal cultures were examined adjusted with Adobe Photoshop. in this study. Vegetative cells of the P. oceanica examined here were collected by S. Matsumoto at Yumigahama DNA methods Beach, Minamiizu, Shizuoka Prefecture, Japan, on 20 May c. 2005, attached to Cladophora sp., and by B.K. Petkus at Samples of 500 ml of culture were filtered through 3-mm- Horseneck State Beach, Westport, Massachusetts, USA, on pore-diameter membrane filters (Millipore). Filters were August 2006, from bottom sand. For morphological immersed in 500 ml DNA extraction buffer containing 2% comparison, S. unipunctata, the generitype of the genus (w/v) CTAB, 1.4 M NaCl, 20 mM EDTA, 100 mM Tris- Striatella, was collected by L.K. Medlin from Banyuls sur HCl, pH 8, 0.2% (w/v) PVP, 0.01% (w/v) SDS and 0.2% b- Mer, France, on 13 February 2005. Single cells were mercaptoethanol. Immersed filters were incubated at 65uC isolated from the American and French samples to obtain for 5 min, vortexed for a few seconds and then discarded. clonal cultures. Cultures were maintained in IMR medium Subsequently, the buffer was cooled briefly on ice. DNA was extracted with an equal volume of chloroform–isoamyl (Eppley et al. 1967) at 15uC under cool-white fluorescent light on a 14 : 10-h (L : D) photoperiod at a photon flux alcohol (24 : 1 [v/v]) and centrifuged in a tabletop density of 30–40 mmol photons m22 s21. A coverslip was Eppendorf microfuge (Eppendorf) at maximum speed placed on the bottom of the culture vessel to be colonized (14,000 rpm) for 10 min. The aqueous phase was collected, with cells producing auxospores. Both strains examined in re-extracted with chloroform–isoamyl alcohol and centri- this study, P. oceanica s0384 and S. unipunctata s0208, are fuged as described previously. Next, the aqueous phase was currently available on request to the first author but may mixed thoroughly with 0.8 volumes of ice-cold 100% not survive long-term in culture (cf. Chepurnov et al. 2004). isopropanol, left on ice for 5 min and subsequently centrifuged in a precooled Eppendorf microfuge at Microscopy maximum speed for 15 min. DNA pellets were washed in 500 ml 70% (v/v) ethanol, centrifuged for 6 min and then An Axioplan (Zeiss) light microscope (LM) with bright allowed to air-dry after decanting off the ethanol. DNA field, differential interference contrast (DIC) or phase pellets were dissolved overnight in 100 ml water. The contrast optics was used to observe living cells and cleaned quantity and quality of DNA were examined by agarose frustules. To photograph live specimens attached to the gel electrophoresis against known standards. bottom of the culture vessel, a Zeiss Axiovert 35 inverted The targeted marker sequence comprised the 18S rDNA microscope was used, equipped with an AxioCam MRc within the nuclear rDNA cistron. The marker was PCR- digital camera. To remove organic material from the amplified in 25-ml volumes containing 10 ng DNA, 1 mM frustule, samples were treated as follows (modified from dNTPs, 0.5 mM of forward primer, 0.5 mM of reverse Nagumo & Kobayashi 1990): (1) the sample was centrifuged primer, 13 Roche diagnostics PCR reaction buffer (Roche to make a pellet and the supernatant discarded; (2) the pellet Diagnostics) and 1 unit Taq DNA polymerase (Roche). The was resuspended in distilled water, and steps 1 and 2 were PCR cycling comprised an initial 4-min heating step at then repeated several times to remove salts; (3) to remove 94uC, followed by 35 cycles of 94uC for 2 min, 56uC for organic matter, the pellet was suspended (using a vortex 4 min and 72uC for 2 min and a final extension at 72uC for mixer) in an equal volume of Drano Power-Gel (Johnson 10 min. PCR products were generated using the forward Wax), a strong domestic drain cleaner; (4) the suspension primer A and a reverse primer B (Medlin et al. 1988) Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 373 without the polylinkers. The quantity and length of using different random starting MP trees (one round of products were examined by agarose gel electrophoresis taxon addition) and the rapid hill-climbing algorithm (i.e. against known standards. Excess primers and dNTPs were option -f d in RAxML). Bootstrap values were obtained by removed from PCR product using the QIAQuick purifica- 100 replications with the GTRCAT model. tion kit (QIAGEN), following the manufacturer’s instruc- The Message Passing Interface (MPI) version of tions. The cleaned PCR products were then electrophoresed MrBayes 3.1.2 (Huelsenbeck & Ronquist 2001; Ronquist on an ABI 3100 Avant sequencer (Applied Biosystems) & Huelsenbeck 2003; Altekar et al. 2004) was used for using Big Dye Terminator v. 3.1 sequencing chemistry Bayesian analyses with the GTR + I + G model to estimate (Applied Biosystems) with the sequencing primers specified the posterior probability distribution using Metropolis- by Elwood et al. (1985). Coupled Markov chain Monte Carlo (MCMCMC) (Ron- quist & Huelsenbeck 2003). The MCMCMC from a Data analyses random starting tree was used in this analysis with two independent runs and one cold and three heated chains. The obtained 18S rDNA sequences were aligned with The Bayesian analyses were run for 20 million generations publicly available sequences retrieved from GenBank each with trees sampled every 100th generation. To increase (Table 1), first using ClustalX (Thompson et al. 1997) the probability of chain convergence, we sampled trees after and then refined by referring to the secondary structure the standard deviation values of the two runs dipped below model of the 18S rRNA at the database of the structure of 0.01 to calculate the posterior probabilities (i.e. after rRNA (Van de Peer et al. 1998). There is extreme length 8,300,000 generations). The remaining phylogenies were variation in some rRNAs (e.g. Gillespie et al. 2005), and discarded as burn-in. replication slippage often leads to convergence on similar primary and secondary structures (Hancock & Vogler 2000; Terminology Shull et al. 2001). Homology assessment in such regions was difficult or impossible, so that the highly variable regions Terminology follows Anonymous (1975) and (particular- (most peripheral regions of the 18S rRNA secondary ly for auxospore structures) Round et al. (1990). structure) were removed from the alignment using BioEdit Molecular phylogenetic studies of diatoms have revealed 7.0.2 (Hall 1999) by referring to the variability map of that historical diatom classifications do not reflect a Saccharomyces cerevisiae (Van de Peer et al. 1993), resulting natural system and that araphid pennate diatoms are in 1713 nucleotides in the data set. paraphyletic in most gene phylogenies, for example, using The data set consisted of 181 OTUs including the closest nuclear 18S ribosomal DNA (rDNA) and plastid 16S relatives of the diatoms Bolidomonas mediterranea Guillou rDNA (Medlin & Kaczmarska 2004). Nevertheless, we & Chre´tiennot-Dinet and B. pacifica Guillou & Chre´tien- use the terms araphid and centric here because they refer to not-Dinet (Guillou et al. 1999) as outgroups. The alignment key morphological features or their absence. In this paper, examined in this study is available at TreeBASE (SN3793). the term araphid pennate diatom follows the traditional To determine which model of sequence evolution best fits definition, that is, a diatom that has an elongate valve with a the data, hierarchical likelihood ratio tests and the Akaike central or slightly lateral sternum, apical pore fields and information criterion were performed using Modeltest 3.7 often also apical rimoportulae but that lacks a raphe slit. We (Posada & Crandall 1998), and both tests selected the GTR do not imply that this corresponds to a mono- (holo-) + I + G model. This model had the following parameters: phyletic group or that it should be accorded any taxonomic base frequencies 5 A: 0.2685, C: 0.1643, G: 0.2539 and T: status. 0.3133; substitution rates were A–C 5 1.2232, A–G 5 3.1535, A–T 5 1.2675, C–G 5 1.4955, C–T 5 5.5072 and G–T 5 1.0000; the proportion of invariant sites was 0.2825; RESULTS and among-site rate heterogeneity was described by a gamma distribution with a shape parameter of 0.6058. Pseudostriatella S. Sato, Mann & Medlin gen. nov. Phylogenies were reconstructed with PAUP v. 4.0b10 (Swofford 2002) using neighbour joining (NJ) of likewise- Figs 1–53 constrained pairwise maximum likelihood (ML) distances. Cellulae rectangulares in aspectu cincturae angulis rotundatis, angulo unico per stipitem muci ad substratum adhaerentes. Chlor- Nodal support was estimated using NJ bootstrap analyses oplasti c. 10 dispersi et cellulam complentes. Taeniae cincturae using the same settings (1000 replicates). numerosae apertae septo conspicuo poros aliquot praebenti. Valvae Maximum parsimony (MP) tree searches were done with lanceolatae, ocello ad utrumque polum, fronte in sectione transversali the ‘new technology’ search algorithm implemented in the arcuata, sine limbo distincto. Striae irregulares in LM non Willi Hennig Society edition of TNT 1.1 (http://www.zmuc. manifestae. Sternum typicum nullum sed area hyalina secus axem longam adest. Areolae per clavulas occlusae, igitur aperturis dk/public/phylogeny/TNT). One hundred random addition dendriticis instructae. Rimoportulae multae dispersae, forma interna sequence replicates were performed with default values. variabili. Nonparametric bootstrap analyses were done 1000 times Cells attached to the substratum by a mucilage stalk at one with the ‘traditional’ search algorithm in TNT. corner of the frustule, rectangular in girdle view, with rounded Maximum likeli hood analyses were performed by corners. Plastids c. 10 per cell, scattered and filling the cell. Copulae numerous open hoops, each with a conspicuous septum containing RAxML-VI-HPC, v. 2.2.3 (Stamatakis et al. 2005) with several pores. Valve lanceolate, with an apical pore field (ocellus) at the GTRMIX model. The analyses were performed 100 each pole. Valve face arched, without a distinct mantle. Striae times to find the best topology receiving the best likelihood irregular, unresolved with LM. Sternum apparently absent, but a 374 Phycologia, Vol. 47 (4), 2008

Table 1. List of taxon and GenBank accession numbers for 18S Table 1. Continued rDNA sequences used in this study. Accession Accession Taxon no. Taxon no. Stephanopyxis cf. broschii M87330 Aulacoseira ambigua (Grunow) Simonsen X85404 Thalassiosira eccentrica (Ehrenberg) Cleve X85396 Aulacoseira baicalensis (Meyer) Simonsen AJ535185 Thalassiosira guillardii Hasle AF374478 Aulacoseira baicalensis (Meyer) Simonsen AJ535186 Thalassiosira oceanica Hasle AF374479 Aulacoseira baicalensis (Meyer) Simonsen AY121821 Thalassiosira pseudonana Hasle et Heimdal AJ535169 Aulacoseira distans (Ehrenberg) Simonsen X85403 Thalassiosira pseudonana Hasle et Heimdal AF374481 Aulacoseira islandica (Mu¨ller) Simonsen AJ535183 Thalassiosira rotula Meunier AF374480 Aulacoseira islandica (Mu¨ller) Simonsen AY121820 Thalassiosira rotula Meunier AF462058 Aulacoseira nyassensis (Mu¨ller) Simonsen AJ535187 Thalassiosira rotula Meunier AF462059 Aulacoseira nyassensis (Mu¨ller) Simonsen AY121819 Thalassiosira rotula Meunier X85397 Aulacoseira skvortzowii Edlund, Stoermer et Taylor AJ535184 Thalassiosira weissflogii (Grunow) Fryxell et Hasle AF374477 Aulacoseira subarctica (Mu¨ller) Haworth AY121818 Thalassiosira weissflogii (Grunow) Fryxell et Hasle AJ535170 Actinocyclus curvatulus Janisch X85401 Thalassiosira sp. AJ535171 Actinoptychus seniarius (Ehrenberg) He´ribaud AJ535182 Toxarium undulatum Bailey AF525668 Bellerochea malleus (Brightwell) van Heurck AF525671 Asterionella formosa Hassall AF525657 Biddulphiopsis titiana (Grunow) von Stosch et Asterionellopsis glacialis (Castracane) Round X77701 Simonsen AF525669 Asterionellopsis glacialis (Castracane) Round AY216904 Chaetoceros curvisetus Cleve AY229895 Asteroplanus karianus1 (Grunow in Cleve et Grunow) Chaetoceros debilis Cleve AY229896 Gardner et Crawford Y10568 Chaetoceros didymus Ehrenberg X85392 Cyclophora tenuis Castracane AJ535142 Chaetoceros gracilis Schu¨tt AY229897 Diatoma hyemalis (Roth) Heiberg AB085829 Chaetoceros rostratus Lauder X85391 Diatoma tenue Agardh AJ535143 Chaetoceros sp. AF145226 Fragilaria crotonensis Kitton AF525662 Chaetoceros sp. AJ535167 Fragilariforma virescens (Ralfs) Williams et Round AJ535137 Chaetoceros sp. X85390 Grammatophora gibberula Ku¨tzing AF525656 Corethron criophilum Castracane X85400 Grammatophora oceanica Ehrenberg AF525655 Corethron inerme Karsten AJ535180 Grammatophora marina (Lyngbye) Ku¨tzing AY216906 Corethron hystrix Hensen AJ535179 Grammonema striatula Agardh1 X77704 Coscinodiscus radiatus Ehrenberg X77705 Grammonema cf. islandica1 AJ535190 Cyclotella meneghiniana Ku¨tzing AJ535172 Grammonema sp.1 AJ535141 Cyclotella meneghiniana Ku¨tzing AY496206 Hyalosira delicatula Ku¨tzing AF525654 Cyclotella meneghiniana Ku¨tzing AY496207 Licmophora juergensii Agardh AF525661 Cyclotella meneghiniana Ku¨tzing AY496210 Nanofrustulum shiloi (Lee, Reimer et McEnery) Round, Cyclotella meneghiniana Ku¨tzing AY496212 Hallsteinsen et Paasche AF525658 Cyclotella cf. scaldensis AY496208 Pseudostriatella oceanica S. Sato, Mann et Medlin AB379680 Cymatosira belgica Grunow X85387 Rhabdonema arcuatum (Agardh) Ku¨tzing AF525660 Detonula confervacea (Cleve) Gran AF525672 Rhaphoneis cf. belgica (Grunow in van Heurck) Ditylum brightwellii (West) Grunow in Van Heurck AY188181 Grunow in van Heurck X77703 Ditylum brightwellii (West) Grunow in Van Heurck AY188182 Staurosira construens Ehrenberg AF525659 Ditylum brightwellii (West) Grunow in Van Heurck X85386 Striatella unipunctata (Lyngbye) Agardh AF525666 Eucampia antarctica (Castracane) Mangin X85389 Synedra sp.2 AJ535138 Guinardia delicatula (Cleve) Hasle AJ535192 Tabularia tabulata (Agardh) Williams et Round AY216907 Guinardia flaccida (Castracane) H. Peragallo AJ535191 Talaroneis posidoniae Kooistra et De Stefano AY216905 Helicotheca tamesis (Schrubsole) Ricard X85385 Thalassionema nitzschioides (Grunow) Hustedt X77702 Lampriscus kittonii Schmidt AF525667 Thalassionema sp. AJ535140 Lauderia borealis Cleve X85399 Synedra ulna Nitzsch AJ535139 Leptocylindrus danicus Cleve AJ535175 Achnanthes bongrainii (M. Peragallo) A. Mann AJ535150 Leptocylindrus minimus Gran AJ535176 Achnanthes sp. AJ535151 Lithodesmium undulatum Ehrenberg Y10569 Amphora montana Krasske AJ243061 Melosira varians Agardh AJ243065 Amphora cf. capitellata AJ535158 Melosira varians Agardh X85402 Amphora cf. proteus AJ535147 Odontella sinensis (Greville) Grunow Y10570 Anomoeoneis sphaerophora (Ehrenberg) Pfitzer AJ535153 Papiliocellulus elegans Hasle, von Stosch et Syvertsen X85388 Bacillaria paxillifer (Mu¨ller) Hendey M87325 Paralia sol (Ehrenberg) Crawford AJ535174 Campylodiscus ralfsii Gregory AJ535162 Planktoniella sol (Wallich) Schu¨tt AJ535173 Cocconeis cf. molesta AJ535148 Pleurosira laevis (Ehrenberg) Compere´ AF525670 Cylindrotheca closterium (Ehrenberg) Reimann et Porosira pseudodenticulata (Hustedt) Jouse´ X85398 Lewin M87326 Proboscia alata (Brightwell) Sundstro¨m AJ535181 Cymbella cymbiformis Agardh AJ535156 Rhizosolenia imbricate Brightwell AJ535178 Encyonema triangulatum Ku¨tzing AJ535157 Rhizosolenia similoides Cleve-Euler J535177 Entomoneis cf. alata AJ535160 Rhizosolenia setigera Brightwell M87329 Eolimna minima (Grunow) Lange-Bertalot AJ243063 Skeletonema costatum (Greville) Cleve X52006 Eolimna subminuscula (Manguin) Moser, Lange- Skeletonema costatum (Greville) Cleve X85395 Bertalot et Metzeltin AJ243064 Skeletonema menzelii Guillard, Carpenter et Reimer AJ535168 Eunotia formica var. sumatrana Hustedt AB085830 Skeletonema menzelii Guillard, Carpenter et Reimer AJ536450 Eunotia monodon var. asiatica Skvortzow AB085831 Skeletonema pseudocostatum Medlin AF462060 Eunotia pectinalis (Dillwyn) Rabenhorst AB085832 Skeletonema pseudocostatum Medlin X85393 Eunotia cf. pectinalis f. minor AJ535146 Skeletonema subsalsum (Cleve-Euler) Bethge AJ535166 Eunotia sp. AJ535145 Skeletonema sp. AJ535165 Fragilariopsis sublineata Hasle AF525665 Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 375

Table 1. Continued ISOTYPE: TNS-AL-53995.

TYPE LOCALITY: Horseneck State Beach, Westport, MA, Accession USA. Taxon no. Gomphonema parvulum Ku¨tzing AJ243062 DISTRIBUTION: Known only from the type locality and Gomphonema pseudoaugur Lange-Bertalot AB085833 Yumigahama Beach, Minamiizu, Shizuoka Prefecture, Japan. Lyrella atlantica (Schmidt) D. G. Mann AJ544659 Navicula cryptocephala var. veneta (Ku¨tzing) Grunow AJ297724 Navicula diserta Hustedt AJ535159 Morphology of vegetative cells Navicula pelliculosa (Bre´bisson ex Ku¨tzing) Hilse AJ544657 Nitzschia apiculata (Gregory) Grunow M87334 In the culture vessel, cells of P. oceanica attached to the Nitzschia frustulum (Ku¨tzing) Grunow AJ535164 bottom, usually by means of a mucilaginous stalk secreted Pinnularia cf. interrupta AJ544658 from one corner, which reached a maximum length of c. Pinnularia sp. AJ535154 20 mm (Fig. 1). Cells occasionally attached to each other to Phaeodactylum tricornutum Bohlin AJ269501 Planothidium lanceolatum (Bre´bisson ex Ku¨tzing) make zigzag chains (not shown), but this was seen only in Round et Bukhtiyarova AJ535189 culture, and chains of over five cells were never found. Pleurosigma sp. AF525664 About 10 lobed plastids were scattered through the cell Pseudogomphonema sp. AF525663 (Fig. 2). A prominent body, likely a pyrenoid, was often Pseudogomphonema sp. AJ535152 visible at the centre of a plastid (Fig. 2, arrow). In girdle Pseudo-nitzschia multiseries (Hasle) Hasle U18241 Pseudo-nitzschia pungens (Grunow ex Cleve) Hasle U18240 view, when the focus was on the surface of the cell, Rossia sp. AJ535144 longitudinal rib-like thickenings were visible on some of the Sellaphora capitata Mann et McDonald AJ535155 girdle bands (Fig. 3). Spots could sometimes be seen on the Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544645 valve mantle (e.g. at arrow, Fig. 3), and tiny projections Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544651 Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544647 were sometimes visible extending inwards from the valve Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544648 face (e.g. at the centre of the valve in Fig. 4); these features Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544649 probably represent rimoportulae (see below). Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544650 The valves were lanceolate with acute ends (Figs 5, 6). Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544652 Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544653 No striae could be resolved in LM with bright field (Fig. 5), Sellaphora pupula (Ku¨tzing) Mereschkowsky AJ544654 DIC (Fig. 6) or phase contrast optics (not shown). The Sellaphora laevissima (Ku¨tzing) D. G. Mann AJ544655 valve length was 16.0–47.8 mm. Auxospore mother cells Sellaphora laevissima (Ku¨tzing) D. G. Mann AJ544656 measured 16.9 6 0.6 mm (mean 6 s, n 5 8), and initial cells Surirella fastuosa var. cuneata (Schmidt) H. Peragallo et M. Peragallo AJ535161 were 41.1 6 3.1 mm(n 5 12). The valve width was 4.4– Thalassiosira antarctica Comber AF374482 5.3 mm. Apical pore fields were recognisable in LM as Undatella sp. AJ535163 hyaline areas at both ends of a valve (Figs 5, 6). Bolidomonas mediterranea Guillou et The frustule had numerous girdle bands (c. 10 per theca: Chrete´innot-Dinet AF123596 Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF123595 Figs 8, 10), each being an incomplete hoop, open at one end Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF167153 (Fig. 7). The closed end of each band bore a septum, which Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF167154 extended inwards by one-sixth to one-eighth of the valve Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF167155 length (Figs 4, 7). The alternation of the septa (Figs 8, 12, Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF167156 32) gave a Striatella-like appearance to the cell in girdle Bolidomonas pacifica Guillou et Chre´teinnot-Dinet AF167157 Convoluta convoluta diatom endosymbiont AY345013 view (Fig. 3). Peridinium foliaceum endosymbiont Y10567 Observations of freeze-dried specimens from field mate- Peridinium balticum endosymbiont Y10566 rial revealed that the surface of the host plant was covered Uncultured diatom AY180014 with bacteria (Figs 9, 10, 13). The long mucilaginous stalk Uncultured diatom AY180015 Uncultured diatom AY180016 (Fig. 9) was secreted from one of the apical pore fields Uncultured diatom AY180020 (Fig. 11), but no secretion occurred from other (Fig. 12). Uncultured eukaryote AY082977 The thickness of the stalk was c. 1 mm (Figs 9, 10, 13). The Uncultured eukaryote AY082992 surface of the mucilaginous stalk was not uniform but Uncultured marine diatom AF290085 comprised many fine strings and thus appeared fibrous 1 Name change since deposit. (Fig. 13). 2 Likely a new genus collected from a marine habitat (Medlin The valve face was smoothly rounded (Fig. 14), lacking et al. 2008a). an abrupt change between it and the mantle. The areolae were irregularly scattered over some parts of the valve, especially towards the poles, but formed parallel striae towards the centre and radiating striae elsewhere (Fig. 14). hyaline area is present along the long axis. Areolae occluded by peg-like structures and therefore with dendritic apertures. Many A clearly defined sternum was absent, but there was an rimoportulae present, of variable form internally. irregular hyaline area along the long axis of the valve (Fig. 14). This hyaline area (1) did not occupy the whole TYPE SPECIES: P. oceanica S. Sato, Mann & Medlin sp. nov. but at most only about one-half of the long axis of the valve Descriptio speciei eadem est ac descriptio generis; valvae 16.0– (Fig. 14), (2) was wider at both ends than in the centre 47.8 mm longae, 4.4–5.3 mm latae. (Fig. 14), and (3) was perforated by small pores in the two HOLOTYPE: BRM Zu6/38. wide end sections (Fig. 15). Apical pore fields were present 376 Phycologia, Vol. 47 (4), 2008

Figs 1–7. Pseudostriatella oceanica: living and cleaned cells (LM). Scale bars 5 100 mm (Fig. 1), 10 mm (Figs 2, 3) or 5 mm (Figs 5–7). Fig. 1. Living cells growing in culture vessel. Fig. 2. Living cell showing multiple plastids. Arrow indicates presumable pyrenoid. Fig. 3. Cleaned frustule, focused on surface to show prominent ribs along the girdle bands, continuous with the septa. The arrow indicates a white spot on the valve that is probably a rimoportula. Fig. 4. Median focus of the frustule in Fig. 3, showing septate girdle bands; arrow indicates probable rimoportula. Fig. 5. Cleaned valve (bright field). Fig. 6. Cleaned valve (DIC). Fig. 7. Single girdle band with septum at closed end. Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 377

Figs 8–13. Pseudostriatella oceanica; intact cells, SEM. Scale bars 5 10 mm (Figs 8, 10), 100 mm (Fig. 9) or 2 mm (Figs 11–13). Fig. 8. Frustule with mucilage stalk secreted at upper right corner. Fig. 9. Colonies on Cladophora sp. Note the cells raised above bacterial community on algal surface by long mucilage stalks. Fig. 10. Side view of frustules just after cell division. Fig. 11. Mucilage stalk secreted from apical pore field. Fig. 12. Free end of valve showing apical pore field secreting no mucilage. Fig. 13. Mucilage stalk attachment to substratum. Note stalk is composed of fine mucilaginous strings. at both ends of the valve. They had slightly thickened rims the Lithodesmiales. Rarely, fused processes were observed and contained small round pores in a strict hexagonal array (Fig. 28). The basal part of each rimoportula always (Fig. 16). This structure conforms to the definition of an overlapped part of an areola (Figs 25–28). Externally, the ocellus (Ross et al. 1979). The areolae were more or less openings of the rimoportulae were undetectable (Figs 14– isodiametric, and each was occluded by two to several pegs 18). leaving a dendritic aperture (Figs 17, 18). The closed end of each girdle band bore a septum On the inner surface of the valve, the hyaline area was (Fig. 29), which was irregularly perforated by many less obvious than on the exterior but still recognizable scattered pores of variable size (Figs 29–31). The pars (Fig. 19). The peg-like occlusions of the areolae were exterior was perforated by simple slit-like areolae arranged slightly sunk below the surface internally (Fig. 22), in short striae (Fig. 32); its margins were plain (Fig. 32, suggesting that these structures were external developments arrow and arrowhead, respectively). The closed end of the (contrast Figs 17, 18). A few of the pores within the band was widened in the pervalvar direction to form a widened ends of the hyaline area penetrated the valve ligula (pointing towards the valve) and a smaller antiligula (Fig. 20). The ocelli were not rimmed internally (Fig. 21), in (pointing away from the valve), which were also perforated contrast to the external appearance (Fig. 16). Approxi- (Fig. 33). Towards the open ends of the band, the septum mately 15–30 rimoportulae were found scattered around became shallower, finally becoming a simple interstria the edge of each valve (Figs 19, 23, 24), all of them having a (Fig. 33). The girdle band areolae were slightly sunken similar size but varying in shape (Figs 25–28). The most internally (Figs 33, 34). common form was ‘C-shaped’, the lips of the process being continuous on one side. This type of process could be either Auxospore structure circular (Fig. 25) or elliptical (Fig. 26). Processes with two entire labiate slits (Fig. 27) were also commonly seen. This Auxosporulation occurred spontaneously in the clonal bilabiate process is not the same as the bilabiate process in culture. Nuclear behaviour was not observed in this study, 378 Phycologia, Vol. 47 (4), 2008

Figs 14–18. Pseudostriatella oceanica valves: external views (SEM). Scale bars 5 5 mm (Fig. 14), 0.5 mm (Figs 15, 16, 18) or 0.2 mm (Fig. 17). Fig. 14. Whole showing central hyaline area (arrow) and irregular striation. Fig. 15. Enlarged view of part marked by asterisk in Fig. 14 showing small simple pores within the end of the hyaline area. Fig. 16. Detail of apical pore field surrounded by plain rim. Fig. 17. Areolae occluded by pegs that vary in shape and number. Fig. 18. Broken valve showing simple nonchambered valve structure. and the earliest stages directly observed were the dehiscence observed at any stage of auxosporulation. All mother cells of the auxospore mother cell, the liberation of the observed in this study were associated with only a single protoplast from its frustule and its bodily movement to a auxospore (Figs 35–38). position beyond the open end of one vacated theca In the earliest stage observed by SEM (Fig. 39), the (Figs 35–37). The auxospore maintained this position organic spherical auxospore did not seem to be covered with subsequently (Figs 37, 38) and must have been physically a mucilage layer or with any siliceous structures. Slightly connected to the empty mother-cell wall, presumably by expanded auxospores (Fig. 40), however, were bordered by mucilage, but the exact nature of the connection could not a plain fringe of material, which probably represented a be established. The young auxospore was more or less delicate mucilaginous envelope (Fig. 41). In auxospores that spherical (Fig. 37). It then expanded at right angles to the had already expanded significantly (Figs 42, 43), a striated pervalvar axis of the gametangium and parallel to its siliceous structure (Fig. 43) could be seen within the longitudinal axis (Fig. 38). No perizonial caps were mucilaginous layer (Fig. 43), and this can probably be Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 379

Figs 19–28. Pseudostriatella oceanica valves: internal views (SEM). Scale bars 5 5 mm (Fig. 19), 0.5 mm (Figs 20, 21), 0.2 mm (Figs 22, 25– 28) or 0.3 mm (Figs 23, 24). Fig. 19. Whole interior. The arrow indicates the hyaline area. Note the many irregularly scattered rimoportulae. Fig. 20. Enlarged view of part marked by asterisk in Fig. 19 showing a few small, simple pores within the hyaline area. Fig. 21. Detail of apical pore field. Fig. 22. Areolae occluded by pegs, which are slightly recessed below the internal valve surface. Figs 23, 24. Broken valve showing irregularly distributed rimoportulae around the valve margin. Figs 25, 26. Circular and elliptical C-shaped rimoportulae. Fig. 27. Normal ‘labiate’ rimoportula. Fig. 28. Compound rimoportula. 380 Phycologia, Vol. 47 (4), 2008

Figs 29–34. Pseudostriatella oceanica girdle (SEM). Scale bars 5 5 mm (Fig. 29), 1 mm (Figs 30, 31) or 2 mm (Figs 32–34). Fig. 29. Single band with a perforated septum. Figs 30, 31. Variation of perforation pattern in septa. Fig. 32. Complete girdle showing interlocking bands. Note the regular striation, except for a hyaline area along the long axis (arrow) and advalvar edge (arrowhead). Fig. 33. Disrupted cingulum showing the outside and inside of an open end. (right) and the outside of a closed end. The plain longitudinal strip is thickened and rib-like. Note that the interstriae region are also rib-like. Fig. 34. Broken copula showing the inside of a closed end. Note that the longitudinal rib becomes more prominent and widens into the septum. regarded as a longitudinal perizonial (LP) band (cf. Fig. 54). very weakly silicified (Figs 45–47) but could be seen to be No transverse perizonial (TP) bands were seen in this stage. finely and irregularly porous (Figs 46, 47, 49). Like the first The LP band comprised a longitudinal rib and a series of LP band seen earlier in expansion, the TP bands also often closely spaced ribs extending out from it at right angles consisted of a longitudinal rib bearing transverse ribs (Fig. 43). At this stage, the body of cell appeared lumpy, (Fig. 47, arrow and arrowhead, respectively), though these which may represent chloroplasts within the auxospore or a were very feebly developed. The ends of the transverse partial covering of silica scales (see below). elements of the TP bands sometimes bore a fringe (Fig. 47). When the expansion was complete, an initial valve was The series of TP bands and LP bands covered the produced within the auxospore (Fig. 44). By then, the auxospore (Figs 48, 54). We will refer to the side closest to auxospore could be seen to possess not only longitudinal the theca of the auxospore mother cell as ‘ventral’, and it but also transverse bands (Figs 45, 46). The TP bands were was on this side that the LP bands lay. There were several Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 381

Figs 35–38. Pseudostriatella oceanica: clonal auxosporulation (LM). Scale bars 5 10 mm. Figs 35, 36. Young auxospores being liberated from their mother cells. Fig. 37. Contracted 6 spherical auxospore. Fig. 38. Mature auxospore containing initial cell, lying slightly oblique to the object plane.

Figs 39–43. Pseudostriatella oceanica: early stages of auxosporulation (SEM). Scale bars 5 10 mm (Figs 39, 40, 42) or 5 mm (Figs 41, 43). Fig. 39. Spherical auxospore. No covering is visible. Fig. 40. Slightly expanded auxospore. Fig. 41. Enlarged view of auxospore of Fig. 40. Arrow indicates mucilaginous layer covering auxospore. Fig. 42. Expanding auxospore. Fig. 43. Enlarged view of auxospore of Fig. 42. Arrow indicates mucilaginous layer. Arrowhead and double arrowhead indicate longitudinal and transverse ribs of a LP band, which has probably been bent during specimen preparation (cf. Fig. 54). Note that a transverse perizonium is absent. 382 Phycologia, Vol. 47 (4), 2008

Figs 44–47. Pseudostriatella oceanica: fully expanded auxospore containing an immature initial epivalve (SEM). Scale bars 5 10 mm (Fig. 44), 5 mm(Fig.45)or2mm (Figs 46, 47). Fig. 44. Whole auxospore still associated with auxospore mother cell. Fig. 45. Enlarged view of auxospore: the initial valve is detectable via its larger, coarser areolae, visible along the midline of the collapsed cell. The auxospore is covered by transverse perizonial bands (TP) dorsally and longitudinal perizonial bands (LP) ventrally. Fig. 46. Enlargement (at black asterisk in Fig. 45), showing the structural differences between the transverse (TP) and longitudinal perizonial bands (LP). Fig. 47. Enlargement (at white asterisk in Fig. 45), showing the delicate TP bands. No regular striae exist. The arrowhead and double arrowhead indicate the longitudinal and transverse ribs of a TP band, respectively; the arrow indicates a fringe.

TP bands (Fig. 49), all closed hoops except for the primary and difficult to detect under SEM, unless a strongly band, which was a simple strap passing from one side of the contrasted image was produced. They varied in shape, but auxospore to the other (Fig. 54). The closed hoops, which in the most frequent type there was a central pore field we will refer to as secondary bands, did not girdle the ringed by a prominent annulus bearing a delicate fringe of auxospore fully. Instead, each was divided into two broader fine radiating fimbriae (Fig. 50, arrow). Much simpler segments on the dorsal side of the auxospore, connected by scales were also found (Fig. 50, arrowhead), lacking the a narrower strip along the ventral side (Figs 49, 54). prominent annulus and fringe; they were sometimes fused The primary band and the adjacent secondary TP band to each other. (Fig. 49: ‘1st’ and ‘2nd’, respectively) were so delicate and When initial cell formation was complete, the series of TP closely associated that the boundary between them was bands was released from the initial cell (Fig. 51, arrow). indistinct, and they may even have been fused together. The There were at least three LP bands. The widest lay furthest more distal TP bands (‘3rd’ and ‘4th’ in Figs 49, 54) towards the ventral side (Fig. 52) and was always flanked in appeared to be slightly more robust than the first two bands our images by two narrower bands (Figs 52, 53; see also and bore distinct longitudinal ribs. The fringed margins of Fig. 54). The transverse ribs of the LP bands were not the bands never interlocked with each other (except perhaps straight, plain straps but sinuous or branched or even fused between the primary band and its neighbour) but over- (Fig. 53). lapped from the centre outwards, the presumably older bands being external to the younger ones (Fig. 49). Phylogeny Rimoportulae were detected on the initial epivalve (Fig. 49). In fully expanded auxospores, scales were found Highly variable regions were excluded from the 18S rDNA on the ventral side (Fig. 50). They were thin and delicate alignment. The analysis used 1713 aligned positions, and Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 383 for this, P. oceanica differed from its closest relative, S. 1990; our unpublished observations), but in Striatella they unipunctata, in 104 substitutions and 28 indels. are not associated with each other. Instead, the sternum A Bayesian tree inferred from 18S rDNA sequences of ends some distance short of the apical pore field, and the 174 diatoms and seven Bolidophyceae (Table 1) confirmed rimoportula lies within a small area of apically orientated paraphyly for the araphid diatoms within a robust clade of striae. The rimoportulae lie within the striae in Pseudos- pennate diatoms (Fig. 55). The same result was also triatella also but are not restricted to the cell apex. obtained with NJ, MP and ML analyses (topologies not shown). Within the pennates, the Asterionellopsis Round– Comparison of Pseudostriatella and Striatella Asteroplanus Gardner & Crawford–Talaroneis Kooistra & De Stefano clade and Rhaphoneis cf. belgica emerged first, There are many morphological and ecological similarities and then a clade containing all other sequenced araphid between P. oceanica and S. unipunctata, such as the pennates and a clade of raphid diatoms diverged (Fig. 55; numerous copulae and their areolation and prominent the outgroup centric diatoms and bolidomonads have been septa, the attachment of cells to substrata in the marine omitted for clarity). Pseudostriatella oceanica formed a littoral and sublittoral by a long mucilaginous stalk and the robust monophyly with S. unipunctata, this in turn production of this stalk via a rimmed apical pore field emerging from within raphid diatoms, being sister to the (ocellus). Because of the morphological similarities of P. genus Achnanthes Bory (MP) or Eunotia Ehrenberg (NJ, oceanica and S. unipunctata, especially with LM, it is quite ML and BI; only the Bayesian tree is shown in Fig. 55). possible that the species has been misidentified as S. unipunctata in the past. On the other hand, there are also many differences, which we regard as sufficient to differentiate these taxa at the rank of genus. The most DISCUSSION striking features are the unusual striation, prominent hyaline area, pegged areolae, multiple marginal rimoportu- Taxonomic comment on the order Striatellales and the lae and perforated septum. Furthermore, if living specimens family Striatellaceae are obtainable, it is easy to identify them because the Round et al. (1990) proposed that the genus Striatella plastids of S. unipunctata are unmistakable because of their Agardh should be regarded as a monospecific genus rod shape and radial arrangement (Fig. 56). With care, because the other (rarely reported) species assigned to it cleaned material of the genera can also be separated in LM. (Van Landingham 1978) differ from the type species S. Thus, in S. unipunctata, each corner of the frustule is sharply unipunctata. We have encountered and isolated only S. truncated (Fig. 56) because of the sunken apical pore fields unipunctata during this study, and the plastids and fine (see Round et al. 1990, p. 432, figs d, e); whereas, rounded structure of the other species are unknown. Two of the 12 corners occur in P. oceanica (Figs 2–4). Again, in valve species considered to be valid by Van Landingham (1978) views of S. unipunctata, the sternum is prominent and runs have been transferred to Hyalosira Ku¨tzing by Navarro & almost the whole length of the valve, the striae are regularly Williams (1991). There are thus 10 species currently in the arranged (with staggered areolae giving a pattern of genus Striatella. transversely orientated diamonds) and can be observed Striatella is the nominate genus of the order Striatellales, (Fig. 57) and the apical rimoportulae are clearly visible; which was established by Round et al. (1990) and contains none of these features exist in P. oceanica (Figs 5, 6). the single family Striatellaceae (Ku¨tzing 1844). In turn, Pseudostriatella oceanica is smaller than S. unipunctata. Round considered the Striatellaceae as comprising three The range observed for P. oceanica (16.0–47.8 mm) is marine benthic genera: Striatella, Hyalosira and Gramma- probably close to the maximum for the species because we tophora Ehrenberg. Judging by the description given by observed both auxospore mother cells and initial cells. It is Round et al. (1990, p. 655) for the Striatellales, these three possible that smaller cells may be formed on occasion genera were linked because they have a narrow or indistinct because cells of other species sometimes continue to divide sternum, well-differentiated apical pore fields (in which the after the minimum threshold for sexual reproduction has pores are in a strict hexagonal array) and porous septate been passed (Geitler 1932). Furthermore, the sizes of the girdle bands. They also have a rimoportula at each apex. initial cells can sometimes depend on the sizes of the Molecular phylogenies show, however, that although gametangia or auxospore mother cells (Davidovich 2001), Hyalosira and Grammatophora form a clade, this does not and this seems to be true in S. unipunctata (Chepurnov in contain Striatella, nor is it a close relative of Striatella Roshchin 1994, table 12). However, S. unipunctata can (Fig. 55; see also Sims et al. 2006, fig. 2). These results attain lengths of more than double the maximum seen in P. suggest that the family Striatellaceae and the order oceanica. Because S. unipunctata is widespread in tropical, Striatellales should contain only the nominate genus subtropical and temperate climate zones (Witkowski et al. Striatella and Pseudostriatella; the monophyly of this group 2000), there are many records and measurements of this is strongly supported by 18S rDNA data, and a wider species (e.g. Van Landingham 1978). The widest range, 35– taxonomic revision is currently in preparation using 125 mm, is given by Hustedt (1931). Chepurnov in Roshchin multiple gene markers. With the benefit of hindsight, it is (1994) observed sexual auxosporulation of S. unipunctata in noticeable that Striatella differs from Hyalosira and culture and found that gametangia of 32–42 mm gave rise to Grammatophora in the arrangement of the sternum and initial cells of 107–126 mm; in his illustrations the largest rimoportula. The rimoportula is adjacent or lateral to the auxospore is 154 mm long (measured on his plate 29). In sternum in Hyalosira and Grammatophora (Round et al. monoclonal cultures (which cannot auxosporulate because 384 Phycologia, Vol. 47 (4), 2008 Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 385

Figs 51–53. Pseudostriatella oceanica: initial cell still within auxospore envelope (SEM). Scale bars 5 10 mm (Fig. 51), 5 mm (Fig. 52) or 3 mm (Fig. 53). Fig. 51. Whole initial cell, with auxospore mother cell still attached. The series of TP bands (arrow) appears to be detaching from the initial cell. Fig. 52. Enlarged view of area marked by black asterisk in Fig. 51, showing the seriesofLPbands. Fig. 53. Enlarged view of area marked by white asterisk in Fig. 52. The LP appears to consist of three bands – primary (arrow), secondary (arrowhead) and tertiary (double arrowhead) – but collapse of the auxospore may have hidden two other bands (cf. Fig. 54).

S. unipunctata is dioecious), some cells continued to divide maintained by selection, but their significance is unknown until they were 22 mm long before dying. Overall, therefore, because the function of the rimoportula is unclear. In some although the size ranges of P. oceanica and S. unipunctata cases, it has been shown that diatoms secrete mucilage do overlap, they differ enough that valve length can help to through the rimoportula for movement, as in Actinocyclus distinguish the species in LM. Ehrenberg (Medlin et al. 1986) and Odontella Agardh (Pickett-Heaps et al. 1986), or adhesion, as in Melosira Rimoportula function Agardh (Crawford 1975) and Aulacodiscus Ehrenberg (Sims & Holmes 1983, p. 270). Schmid (1994) has suggested that In araphid diatoms, there is some variation in the the internal part of the rimoportula is used as a cytological distribution of the rimoportulae; although, they are most anchor for the nucleus during interphase and new valve often located along the long axis, mostly near one or both formation, and recently Ku¨hn & Brownlee (2005) have ends of the sternum. Normally, too, each rimoportula has provided evidence that the rimoportula is a site for its own special opening externally, which is separate and endocytosis and therefore involved in membrane recycling. different from the external openings of the areolae. The It is quite possible that rimoportulae serve multiple roles in consistency of these features suggests that they are diatoms (Medlin et al. 1986). In P. oceanica, the rimopor-

r

Figs 48–50. Pseudostriatella oceanica: final stage of auxosporulation (SEM). Scale bars 5 10 mm (Fig. 48) and 5 mm (Fig. 49) and 2 mm (Fig. 50). Fig. 48. Auxospore containing a mature initial epivalve. Fig. 49. Enlarged view of middle of auxospore of Fig. 48, showing that the initial valve is covered by TP and LP (bottom right) bands. TP bands are numbered from primary (1st) to fourth (4th). Note fuzzy border of 1st and 2nd bands. Bands 1 and 2 do not have rib thickenings, whereas bands 3 and 4 bands do (arrow and double arrowhead, respectively). The edge of band 4 overlap onto band 5 (double arrowhead). Triple arrowheads indicate rimoportulae at internal initial valve. Fig. 50. Enlargement of area marked by asterisk in Fig. 48, showing scales on the ventral side of the auxospore. Two types are present, with (arrow) and without (arrowhead) an annulus. 386 Phycologia, Vol. 47 (4), 2008

excluded because the two taxa also share a very long node with high statistical support. Preliminary analyses using several gene markers also show that monophyly of the clade containing the two genera is robust (S. Sato, unpublished observations). Many phylogenetic studies of diatoms made using 18S rDNA have revealed that the araphid pennate diatoms are paraphyletic. They divide into two groups: (1) a relatively small clade of marine diatoms containing the Rhaphonei- daceae, Plagiogrammaceae, Asterionellopsis and Asteropla- nus and (2) a larger, ‘core’ group (grade) containing the rest of the araphid diatoms that is the sister group to the raphid diatoms (e.g. Medlin & Kaczmarska 2004; Alverson et al. 2006; Sims et al. 2006). This relationship was recovered in the present analysis. Some features of our tree, such as the sister relationship between the P. oceanica–S. unipunctata clade and the raphid genus Eunotia, have high support but are frankly implausible because of morphological and Fig. 54. Pseudostriatella oceanica: plan diagram and section (right) reproductive evidence. For example, the pattern of aux- of ‘perizonium’ and initial cell. osporulation in Striatella (cis anisogamy coupled with expansion of the auxospore at the mouth of the female tulae have no external openings of their own and connect to gametangium and at right angles to it: Chepurnov in the outside instead through part of an areola (Figs 25–28). Roshchin 1994) is not shared by Eunotia (Mann et al. 2003) This, together with their scattered distribution on the valve, or with any other raphid diatoms (Round et al. 1990) but makes it unlikely that the rimoportulae function in does agree well, though not perfectly, with auxosporulation movement in P. oceanica or in the production of robust in Rhabdonema Ku¨tzing and Grammatophora (von Stosch structured mucilage for adhesion. Altogether, the charac- 1962; Sato et al. 2008a). There are also no vestiges of a teristics of the rimoportulae in P. oceanica suggest relaxed raphe in Pseudostriatella (contrast Cocconeis Ehrenberg functional constraints, relative to other araphid pennates. ‘pseudoraphe’ valves, Semiorbis R. Patrick and some On the other hand, the rimoportulae have not been lost Asterionella-like diatoms; Mann 1982a; Round et al. 1990; altogether in P. oceanica, in contrast to members of the Kociolek & Rhode 1998). The poorly supported relation- Plagiogrammaceae (including Talaroneis, Dimeregramma ship to the raphid diatoms probably results from well- Ralfs and Plagiogramma Paddock) and other genera, such known analytical artifacts, such as taxon sampling or as Staurosira (Ehrenberg) Williams & Round, Nanofrustu- substitution bias: the long branches seen in P. oceanica–S. lum Round, Hallsteinsen & Paasche, Opephora Petit, unipunctata clade suggest an accelerated rate of base Punctastriata Williams & Round, Staurosirella Meresch- substitution, which may make it difficult to reconstruct kowsky, Pseudostaurosiropsis Morales and Pseudostauro- the phylogeny correctly. sira Williams & Round (Round et al. 1990, 1999; Morales Indeed, 18S rDNA analyses undertaken so far have 2001, 2005; Kooistra et al. 2004). Among these rimopor- placed S. unipunctata in various phylogenetic positions. tula-lacking diatoms, few seem to be able to grow as Some put the genus at the root of the raphid diatoms epiphytes – possibly only Talaroneis (Kooistra et al. 2004); (Kooistra et al. 2003a, b, 2004). Given the hypothesis of the rest grow attached to rocks or sand grains or live Hasle (1974) that the rimoportula might be the predecessor planktonically. Possession of rimoportulae may therefore of the raphe, Kooistra et al. (2003a) implied that the slightly be important in araphid pennates for attachment to plants. elongated external opening of the rimoportula in Striatella illustrates how the raphe could have arisen, by elongation Phylogeny towards the centre of the valve, creating two slits splitting the sternum. By contrast, in Medlin & Kaczmarska’s (2004) The 18S rDNA phylogeny gave strong support not only to analyses, the sister to Striatella is Staurosira construens the monophyly of the P. oceanica–S. unipunctata clade Ehrenberg, which has no morphological features in (Bootstrap supports in NJ and ML analyses 5 100; common with Striatella and Pseudostriatella beyond its Bayesian posterior probability 5 1.0) but also to the elongate shape (Round et al. 1990). In an 18S rDNA tree establishment of a new genus for P. oceanica because of the using almost the same data set of Medlin & Kaczmarska long branches connecting both species. The high divergence (2004) but constructed by direct optimization (DO), a between these taxa (104 substitutions and 28 indels) heuristic maximum parsimony algorithm, Striatella is sister contrasts, for example, with the shorter branch lengths to the marine araphid genus Licmophora Agardh (Sorhan- within Grammatophora Ehrenberg and Eunotia (see nus 2004). Finally, the genus has appeared within the Fig. 55). We accept, of course, that there is no absolute raphid diatoms, as sister to an Anomoeoneis Pfitzer– standard for the amount of sequence difference that justifies Cymbella Agardh clade (Medlin et al. 2000). None of these generic status. Although P. oceanica and S. unipunctata lie relationships are robust (i.e. they receive low bootstrap at the ends of long branches, the possibility that the tree has support or Bayesian posterior probabilities; statistical been distorted by long-branch artifacts can probably be support data are not available in Sorhannus 2004). Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 387

Fig. 55. Molecular phylogeny of araphid pennate diatoms inferred from 18S rDNA sequence using 1713 aligned positions. The tree shown resulted from Bayesian inference using a GTR + I + G model. Outgroup bolidomonads and centric diatoms were excluded, and a clade comprising raphid diatoms is collapsed into triangle for clarity. Nodal support values greater than 50 (NJ, MP and ML) and 0.50 (BI) are shown. Nodes with strong supports (bootstrap support . 90 in NJ, MP and ML, and posterior probability . 0.95 in BI) are shown as thick lines. 1Name change since deposit; 2likely a new genus collected from a marine habitat (Medlin et al. 2008a); 3annotated as Asterionellopsis kariana in GenBank.

One recent result from a Bayesian 18S rDNA analysis, below) and pattern centre in P. oceanica. The pennate using a doublet model that takes base substitutions in diatoms are usually monophyletic in trees based on a rRNA secondary structure into account, constructed a variety of genes (molecular studies of diatoms are listed by robust Striatella–Rhabdonema clade as one of a radiated Mann & Evans 2007), supporting the idea that the ‘pennate’ group of pennate diatoms (Alverson et al. 2006, fig. 5). Bauplan is a synapomorphy, that is, the possession of a However, the consensus most parsimonious tree discon- single longitudinal rib-like element (sternum) at the centre nected Striatella from Rhabdonema and put the taxa into of the pattern and deposited first during valve formation, polytomy (Alverson et al. 2006, fig. 6). Sims et al. (2006) which subtends sets of transverse ribs on either side (e.g. used a huge data set that placed Striatella at the root of the Round et al. 1990, p. 31). Some diatoms previously ‘core’ araphid + raphid clade with high support. In the ML regarded as ‘pennates’, such as Toxarium, Ardissonea and tree presented by Sorhannus (2007), Striatella also diverges Climacosphenia, which have a different kind of pattern at the root of the core araphid clade but with low bootstrap centre (Mann 1984), have been shown to belong outside the support. pennate clade (Kooistra et al. 2003b; Alverson et al. 2006; We conclude, therefore, that it is probably impossible at Medlin et al. 2008b). However, the pattern centre in P. this time to obtain a fully resolved phylogeny resolving the oceanica is unlike anything found previously in pennate correct phylogenetic placement of the P. oceanica–S. diatoms partly because it is a wide unthickened hyaline area unipunctata clade, and it may remain impossible when 18S but more importantly because its wider terminal sections rDNA sequences are used in a single gene phylogeny. It is contain pores. In fact, the hyaline area resembles a highly particularly important to establish the position of the clade elongate annulus – a more extreme version of the elongate because of the unusual structure of the auxospore (see annuli seen in some Odontella (e.g. Pickett-Heaps et al. 388 Phycologia, Vol. 47 (4), 2008

diatoms (e.g. von Stosch 1962, 1982; Crawford 1974; Kobayashi et al. 2001; Schmid & Crawford 2001), the scales varied in shape within a single auxospore. Some scales had an annulus and were morphologically similar to those of centric diatoms (Round et al. 1990). In P. oceanica, the auxospores never had a complete covering of scales. The few scales present were restricted to the ventral side in nearly mature or mature examples. A ventral distribution is also present in fully developed auxospores of the medio- phycean centric diatom Chaetoceros didymum Ehrenberg (von Stosch 1982, fig. 2), although here the scales can also be detected from the earliest stages (von Stosch et al. 1973; von Stosch 1982). Some details of perizonial structure were obscured by collapse of the auxospore during air drying. However, the Figs 56, 57. Striatella unipunctata (LM). Scale bars 5 10 mm. widest LP band was always located at the most ventral end Fig. 56. Living cell showing distinctive plastids. Arrowhead (Fig. 52), and it was associated with two additional bands. indicates truncated corner of cell. We therefore infer that the widest band is the primary band Fig. 57. Cleaned valve taken with phase contrast optics. Arrow and that the additional bands flanking it are a secondary indicates rimoportula, arrowhead indicates sternum. and a tertiary LP band (Fig. 54). We believe, however, that some of the LP bands were hidden by folding of the 1990, fig. 40e) and Attheya species (Crawford et al. 1994). auxospore, which seems likely because all of the longitu- Until a robust phylogeny is available, it will be unclear dinal perizonia reported so far in pennate diatoms are whether the resemblance between the Pseudostriatella structurally symmetrical (e.g. von Stosch 1962, 1982; Mann pattern centre and an annulus is a symplesiomorphy (i.e. 1982b; Toyoda et al. 2005), even in Amphora (Nagumo Pseudostriatella does not have and has never had a true 2003). There would therefore be five longitudinal bands in sternum) or the result of convergent evolution. P. oceanica (Fig. 54), and a similar arrangement has been found in Tabularia parva (Sato et al. 2008b). Interestingly, Auxosporulation and auxospore fine structure even though the valves of P. oceanica have a poorly expressed and irregular sternum–stria system, the LP bands We were not able to establish how the auxospores arise in have strictly parallel patterning, resembling the striation of monoclonal cultures of P. oceanica because the earliest normal araphid diatom valves. stages were not seen. It is very unlikely that the auxospores The usual structure of the transverse perizonium in developed through allogamous sexual reproduction because pennate diatoms – both araphid and raphid – is that there is we are confident that we would have observed the empty a central primary band with a separate series of secondary frustules of any ‘male’ cells close to the expanding bands on each side (von Stosch 1982; Mann 1982b). The auxospores (Roshchin 1994; Chepurnov et al. 2004). primary band is either a short cylinder wholly encircling the Therefore, we have referred to the auxosporulating cells centre of the auxospore (i.e. it is ‘closed’: e.g. Poulı´cˇkova´& as ‘auxospore mother cells’ rather than as gametangia. Mann 2006), or it is a split ring (an ‘open’ band) with its Further work is needed to determine whether auxosporula- ends almost touching (e.g. Sato et al. 2004). The secondary tion involves meiosis and automictic fusion or whether it is bands are usually open, again with their ends closely apomictic. Nonallogamous formation of auxospores and associated. The TP bands combine to form a cigar-shaped vegetative cell enlargement has been recorded in other perizonium with a narrow ventral suture, beneath which araphid pennates, including Grammatophora (Sato et al. there is often a set of LP bands, again differentiated into a 2008a) and Licmophora (Kumar 1978). central primary band and two short flanking series of The structure of the auxospore in P. oceanica is unlike secondary bands (e.g. Mann 1982b; Mann & Stickle 1993; anything described so far and prompts re-examination of Nagumo 2003; Sato et al. 2004; Toyoda et al. 2005). The the nature of ‘perizonia’ and ‘properizonia’. In its overall function of the perizonium appears to be to support and layout, the auxospore casing of P. oceanica resembles the constrain anisometric expansion of the auxospore (Mann envelopes of Rhabdonema (von Stosch 1962, 1982), 1994). Gephyria (Sato et al. 2004) and Grammatophora (Sato et Many centric diatoms also exhibit anisometric expan- al. 2008a) in that it possesses small more or less isodiametric sion, and again this is apparently controlled through the or slightly elongate scales and also a separate series of formation of band-shaped stiffening elements, which longitudinal and transverse bands. However, there are also together constitute a structure called the ‘properizonium’ significant differences, notably in the structure of the (von Stosch 1982). The key difference between properizonia transverse bands and the spatiotemporal organization of and perizonia identified by von Stosch (von Stosch & auxospore development. Kowallik 1969; von Stosch et al. 1973; von Stosch 1982) is There are very few scales in P. oceanica, compared to that perizonia are independent from the original zygote wall other araphid diatoms (Sato et al. 2004, 2008a, b), and we both structurally and developmentally (the perizonium is found them only on the mature auxospore (although we ‘eine von der urspru¨nglichen Zygotenhu¨lle unabha¨ngige cannot wholly exclude that they were present). As in other Struktur’: von Stosch & Kowallik 1969, p. 469). In Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 389 contrast, properizonia are supposed to be structurally and classification into scaly, properizonial and perizonial developmentally continuous with the scale-containing auxospores (cf. the ‘isometric’, ‘anisometric’ and ‘bilateral’ layers that precede them (the otherwise helpful review by auxospores of Kaczmarska et al. 2001) is perhaps too Kaczmarska et al. 2001 may be misleading in this respect). simple but that the development of shape is generally Thus, von Stosch (1982, p. 146) considered that, for the associated with stiffening of the auxospore wall during evolutionary transition from properizonial casings to expansion by silica bands and hoops. perizonial diatoms like Rhabdonema, ‘the first item necessary would be a developmental and spatial hiatus between scale layers and properizonial band systems’. No ACKNOWLEDGEMENTS other features have been identified that are diagnostic for the perizonium vs the properizonium, but it is generally considered (von Stosch 1982; Round et al. 1990; Kacz- The authors are grateful to Beth K. Petkus for collection of marska et al. 2001; Medlin & Kaczmarska 2004) that living specimen and brought us it from the United States to perizonia are characteristic of pennate diatoms; whereas, Germany with her, Richard M. Crawford for correction of properizonia are restricted to some lineages of multipolar the manuscript and discussion, Stephan Frickenhaus for centric diatoms. Our observations of P. oceanica revealed establishing parallel processing for Bayesian analyses, Paul no clear developmental separation between a primary scale- A. Fryxell for helping to translate the Latin diagnosis, bearing wall and the ‘perizonium’. Scales were rare and Friedel Hinz for technical help for LM and SEM, and produced apparently only towards the end of auxospore Masahiko Idei for allowing us to access his poster for the expansion, and it appears too that longitudinal perizonial 17th International Diatom Symposium. We also thank two elements are produced before the transverse perizonium. In anonymous reviewers for their valuable comments and retrospect, we believe that a similar continuity of develop- suggestions. This study was supported by DAAD for ment may occur also in Gephyria, where we (Sato et al. doctoral research fellowship to Shinya Sato. 2004) detected scales on the inside of the primary TP band. Another curious, ‘transitional’ feature of P. oceanica auxospores concerns the nature of the secondary TP bands. REFERENCES As noted above, the secondary bands on either side of the primary TP band are separate entities in the perizonia of all ALTEKAR G., DWARKADAS S., HUELSENBECK J.P. & RONQUIST F. raphid and araphid pennates studied until now, as they are 2004. Parallel Metropolis-coupled Markov chain Monte Carlo also in a few properizonia (those of Odontella and for Bayesian phylogenetic inference. Bioinformatics 20: 407–415. Biddulphia; von Stosch 1982). Pseudostriatella is quite ALVERSON A.J., CANNONE J.J., GUTELL R.R. & THERIOT E.C. different because the secondary TP elements are continuous 2006. The evolution of elongate shape in diatoms. Journal of Phycology 42: 655–668. from one end of the auxospore to the other around the AMATO A., ORSINI L., D’ALELIO D. & MONTRESOR M. 2005. Life ventral side (Fig. 54). Each is therefore a complex hoop, cycle, size reduction patterns, and ultrastructure of the pennate shaped like the margin of a saddle. Thus, although the planktonic diatom Pseudo-nitzschia delicatissima (Bacillario- expansion of the auxospore is bipolar in P. oceanica, the phyta). Journal of Phycology 41: 542–556. development of the TP itself is unipolar, beginning from the ANONYMOUS. 1975. Proposals for a standardization of diatom terminology and diagnoses. Nova Hedwigia, Beiheft 53: 323–354. strap-like primary band and extending out both laterally CAHOON L.B. 1999. The role of benthic microalgae in neritic (towards the poles) and ventrally. Exactly the same ecosystems. Oceanography and Marine Biology. An Annual unipolar pattern of development occurs in the properizonia Review 37: 47–86. of Chaetoceros, Bacteriastrum, Attheya, Lithodesmium (here CHEPURNOV V.A., MANN D.G., SABBE K. & VYVERMAN W. 2004. the topology is more complex, because most auxospores are Experimental studies on sexual reproduction in diatoms. International Review of Cytology 237: 91–154. triradiate) and Bellerochea (von Stosch 1982) and also in COHN S.A., SPURCK T.P., PICKETT-HEAPS J.D. & EDGAR L.A. Lampriscus (Idei & Nagumo 2002). Thus, it would be as 1989. Perizonium and initial valve formation in the diatom reasonable to regard the auxospore casing of P. oceanica as Navicula cuspidata (Bacillariophyceae). Journal of Phycology 25: a properizonium as it is to describe it as a true perizonium, 15–26. despite the fact that this species clearly belongs phyloge- CRAWFORD R.M. 1974. The auxospore wall of the marine diatom netically to the pennate lineage. The presence of longitu- Melosira nummuloides (Dillw.) C. Ag. and related species. British Phycological Journal 9: 9–20. dinal elements in the Pseudostriatella casing is not CRAWFORD R.M. 1975. The frustule of the initial cells of some conclusive support for interpretation as a perizonium species of the diatom genus Melosira C. Agardh. Nova Hedwigia, because (1) longitudinal elements are present beneath the Beiheft 53: 37–50. ‘transverse’ bands in the triradiate centric Lithodesmium CRAWFORD R.M., GARDNER C. & MEDLIN L.K. 1994. The genus (von Stosch 1982; see also Round et al. 1990) and (2) the Attheya. I. A description of four new taxa, and the transfer of Gonioceros septentrionalis and G. armatus. Diatom Research 9: longitudinal bands of P. oceanica seem to be formed before 27–51. the transverse bands, whereas during perizonium formation DAVIDOVICH N.A. 2001. Species-specific sizes and size range of in raphid diatoms the converse is true. sexual reproduction in diatoms. In: Proceedings of the 16th A simple conclusion can be drawn: there is simply too International Diatom Symposium (Ed. by A. Economou-Amilli), little information from too few taxa to allow detailed University of Athens, Greece. pp. 191–196. ELWOOD H.J., OLSEN G.J. & SOGIN M.L. 1985. The small subunit analysis of the evolution of auxospore structure, and the ribosomal DNA gene sequences from the hypotrichous ciliates distinction between properizonia and perizonia needs to be Oxytricha nova and Stylonichia pustulata. Molecular Biology and re-evaluated. All that can be said at the moment is that Evolution 2: 399–410. 390 Phycologia, Vol. 47 (4), 2008

EPPLEY R.W., HOLMES R.W. & STRICKLAND J.D.H. 1967. Sinking MANN D.G. 1982b. Structure, life history and systematics of rates of the marine phytoplankton measured with a fluoroch- Rhoicosphenia (Bacillariophyta). II. Auxospore formation and rometer. Journal of Experimental Marine Biology and Ecology 1: perizonium structure of Rh. curvata. Journal of Phycology 18: 191–208. 264–274. GEITLER L. 1932. Der Formwechsel der pennaten Diatomeen. MANN D.G. 1984. An ontogenetic approach to diatom systematics. Archiv fu¨r Protistenkunde 78: 1–226. In: Proceedings of the 7th International Diatom Symposium (Ed. GILLESPIE J.J., MCKENNA C.H., YODER M.J., GUTELL R.R., by D.G. Mann), O. Koeltz, Koenigstein, Germany. pp. 113–144. JOHNSTON J.S., KATHIRITHAMBY J. & COGNATO A.I. 2005. MANN G.D. 1994. The origins of shape and form in diatoms: the Assessing the odd secondary structural properties of nuclear interplay between morphogenetic studies and systematics. In: small subunit ribosomal RNA sequences (SSU) of the twisted- Shape and form in plants and fungi (Ed. by D.S. Ingram & A.J. wing parasites (Insecta: Strepsiptera). Insect Molecular Biology Hudson), Academic Press, London. pp. 17–38. 14: 625–643. MANN D.G. & EVANS M.K. 2007. Molecular genetics and the GUILLOU L., CHRE´ TIENNOT-DINET M.-J., MEDLIN L.K., CLAUSTRE neglected art of diatomics. In: Unravelling the algae – the past, H., LOISEAUX-DE GOE¨ R S. & VAULOT D. 1999. Bolidomonas:a present and future of algal molecular systematics (Ed. by J. Brodie new genus with two species belonging to a new algal class, the & J.M. Lewis), CRC Press, Boca Raton, FL, USA. pp. 232–266. Bolidophyceae class. nov. (Heterokonta). Journal of Phycology MANN D.G. & STICKLE A.J. 1993. Life history and systematics of 35: 368–381. Lyrella. Nova Hedwigia, Beiheft 106: 43–70. HALL T.A. 1999. BioEdit: a user-friendly biological sequence MANN D.G., CHEPURNOV V.A. & IDEI M. 2003. Mating system, alignment editor and analysis program for Windows 95/98/NT. sexual reproduction and auxosporulation in the anomalous Nucleic Acids Symposium Series 41: 95–98. raphid diatom Eunotia (Bacillariophyta). Journal of Phycology HANCOCK J.M. & VOGLER A.P. 2000. How slippage-derived 39: 1067–1084. sequences are incorporated into rRNA variable-region secondary MEDLIN L.K. & KACZMARSKA I. 2004. Evolution of the diatoms: structure: implications for phylogeny reconstruction. Molecular V. Morphological and cytological support for the major clades Phylogenetics and Evolution 14: 366–374. and a taxonomic revision. Phycologia 43: 245–270. HASLE G.R. 1974. The ‘mucilage pore’ of pennate diatoms. Nova MEDLIN L.K., CRAWFORD R.M. & ANDERSEN R.A. 1986. Hedwigia, Beiheft 45: 167–194. Histochemical and ultrastructural evidence for the function of HUELSENBECK J.P. & RONQUIST F. 2001. MRBAYES: Bayesian the labiate process in the movement of centric diatoms. British inference of phylogeny. Bioinformatics 17: 754–755. Phycological Journal 21: 297–301. HUSTEDT F. 1931. Die Kieselalgen Deutschlands, O¨ sterreichs und MEDLIN L., ELWOOD H.J., STICKEL S. & SOGIN M.L. 1988. The der Schweiz. In: Dr L Rabenhorsts Kryptogamenflora von ¨ characterization of enzymatically amplified eukaryotic 16S-like Deutschland, Osterreich und der Schweizvol, vol. 7(2:1). Akade- rRNA coding regions. Gene 71: 491–499. mische Verlagsgesellschaft, Leipzig, Germany. pp. 1–176. MEDLIN L.K., KOOISTRA W.H.C.F. & SCHMID A.M.-M. 2000. A IDEI M. & NAGUMO T. 2002. Auxospore structure of the marine review of the evolution of the diatoms – a total approach using diatom genus Lampriscus with triangular/quadrangular forms. molecules, morphology and geology. In: The origin and early In: Abstracts, 17th International Diatom Symposium (Ed. by M. evolution of the diatoms: fossil, molecular and biogeographical Poulin), Ottawa, Canada. 166 pp. approaches (Ed. by A. Witkowski & J. Sieminska), Szafer KACZMARSKA I., BATES S.S., EHRMAN J.M. & LE´ GER C. 2000. Fine Institute of Botany, Polish Academy of Science, Cracow, Poland. structure of the gamete, auxospore and initial cell in the pennate pp. 13–35. diatom Pseudo-nitzschia multiseries (Bacillariophyta). Nova MEDLIN L.K., JUNG I., BAHULIKAR R., MENDGEN K., KROTH P. & Hedwigia 71: 337–357. KOOISTRA W.H.C.F. 2008a. Evolution of the diatoms. VI. KACZMARSKA I., EHRMAN J.M. & BATES S.S. 2001. A review of Assessment of the new genera in the araphids using molecular auxospore structure, ontogeny and diatom phylogeny. In: Pro- ceedings of the 16th International Diatom Symposium (Ed. by A. data. Nova Hedwigia 133: 81–100. Economou-Amilli), University of Athens, Greece. pp. 153–168. MEDLIN L.K., SATO S., MANN D.G. & KOOISTRA W.H.C.F. 2008b. Molecular evidence confirms sister relationship of Ardissonea, KOBAYASHI A., OSADA K., NAGUMO T. & TANAKA J. 2001. An auxospore of Arachnoidiscus ornatus Ehrenberg. In: Proceedings Climacosphenia and Toxarium within the bipolar centric diatoms of the 16th International Diatom Symposium (Ed. by A. (Bacillariophyta, Mediophyceae) and cladistic analyses confirms Economou-Amilli), University of Athens, Greece. pp. 197–204. that extremely elongated shape has arisen twice in the diatoms Journal of Phycology: in press. KOCIOLEK J.P. & RHODE K. 1998. Raphe vestiges in Asterionella species from Madagascar: evidence for a polyphyletic origin of MORALES E. 2001. Morphological studies in selected fragilarioid the araphid diatoms? Cryptogamie: Algologie 19: 57–74. diatoms (Bacillariophyceae) from Connecticut waters (U.S.A.). KOOISTRA W.H.C.F., DE STEFANO M., MANN D.G. & MEDLIN Proceedings of the Academy of Natural Sciences of Philadelphia L.K. 2003a. The phylogeny of the diatoms. Progress in Molecular 151: 39–54. and Subcellular Biology 33: 59–97. MORALES E. 2005. Observations of the morphology of some known KOOISTRA W., DE STEFANO M., MANN D.G., SALMA N. & MEDLIN and new fragilarioid diatoms (Bacillariophyceae) from rivers in L.K. 2003b. Phylogenetic position of Toxarium, a pennate-like the USA. Phycological Research 53: 113–133. lineage within centric diatoms (Bacillariophyceae). Journal of NAGUMO T. 2003. Taxonomic studies of the subgenus Amphora Phycology 39: 185–197. Cleve of the genus Amphora (Bacillariophyceae) in Japan. KOOISTRA W.H.C.F., FORLANI G., STERRENBURG F.A.S. & DE Bibliotheca Diatomologica 49: 1–265. STEFANO M. 2004. Molecular phylogeny and morphology of the NAGUMO T. & KOBAYASHI H. 1990. The bleaching method for marine diatom Talaroneis posidoniae gen. et sp. nov. (Bacillar- gently loosening and cleaning a single diatom frustule. Diatom 5: iophyta) advocate the return of the Plagiogrammaceae to the 45–50. pennate diatoms. Phycologia 43: 58–67. NAVARRO N.J. & WILLIAMS D.M. 1991. Description of Hyalosira KU¨ HN S.F. & BROWNLEE C. 2005. Membrane organisation and tropicalis sp. nov. (Bacillariophyta) with notes on the status of dynamics in the marine diatom Coscinodiscus wailesii (Bacillar- Hyalosira Ku¨tzing and Microtabella Round. Diatom Research 6: iophyceae). Botanica Marina 48: 297–305. 327–336. KUMAR R. 1978. Auxospore formation in species of the marine PICKETT-HEAPS J.D., HILL D.R.A. & WETHERBEE R. 1986. Cellular diatom Licmophora Agardh. Vero¨ffentlichungen des Instituts fu¨r movement in the centric diatom Odontella sinensis. Journal of Meeresforschungen in Bremerhaven 17: 15–20. Phycology 22: 334–339. KU¨ TZING F.T. 1844. Die kieselschaligen Bacillarien oder Diatomeen. PICKETT-HEAPS J.D., SCHMID A.-M.M. & EDGAR L.A. 1990. The Nordhausen, Germany. 152 pp. cell biology of diatom valve formation. Progress in Phycological MANN D.G. 1982a. Structure, life history and systematics of Research 7: 1–168. Rhoicosphenia (Bacillariophyta) I. The vegetative cell of Rh. POSADA D. & CRANDALL K.A. 1998. Modeltest: testing the model curvata. Journal of Phycology 18: 162–176. of DNA substitution. Bioinformatics 14: 817–818. Sato et al.: Pseudostriatella oceanica gen. et sp. nov. 391

POULı´Cˇ KOVA´ A. & MANN D.G. 2006. Sexual reproduction in SULLIVAN M.J. & CURRIN C.A. 2000. Community structure and Navicula cryptocephala (Bacillariophyceae). Journal of Phycology functional dynamics of benthic microalgae in salt marshes. In: 42: 872–886. Concepts and controversies in tidal marsh ecology (Ed. by M.P. POULı´Cˇ KOVA´ A., MAYAMA S., CHEPURNOV V.A. & MANN D.G. Weinstein & D.A. Kreeger), Kluwer Academic Publishers, 2007. Heterothallic auxosporulation, incunabula and perizonium Dordrecht, Germany. pp. 81–106. in Pinnularia (Bacillariophyceae). European Journal of Phycology SWOFFORD D.L. 2002. PAUP*: phylogenetic analysis using 42: 367–390. parsimony (* and other methods), ver. 4.0b10. Sinauer Associ- RONQUIST F. & HUELSENBECK J.P. 2003. MrBayes 3: Bayesian ates, Sunderland, MA, USA. phylogenetic inference under mixed models. Bioinformatics 19: THOMPSON J.D., GIBSON T.J., PLEWNIAK F., JEANMOUGIN F. & 1572–1574. HIGGINS D.G. 1997. The ClustalX windows interface: flexible ROSHCHIN A.M. 1994. Zhiznennye tsikly diatomovykh vodoroslej. strategies for multiple sequence alignment aided by quality Naukova Dumka, Kiev. 170 pp. analysis tools. Nucleic Acids Research 24: 4876–4882. IFFANY ROSS R., COX E.J., KARAYEVA N.I., MANN D.G., PADDOCK T M.A. 2005. Diatom auxospore scales and early stages in diatom frustule morphogenesis: their potential for use in nanotech- T.B.B., SIMONSEN R. & SIMS P.A. 1979. An amended terminol- ogy for the siliceous components of the diatom cell. Nova nology. Journal of Nanoscience and Nanotechnology 5: 131–139. Hedwigia, Beiheft 64: 513–533. TOYODA K., IDEI M., NAGUMO T. & TANAKA J. 2005. Fine- structure of frustule, perizonium and initial valve of Achnanthes ROUND F.E., CRAWFORD R.M. & MANN D.G. 1990. The diatoms: yaquinensis McIntire and Reimer (Bacillariophyceae). European biology and morphology of the genera. Cambridge University Journal of Phycology 40: 269–279. Press, Cambridge, UK. 747 pp. TOYODA K., WILLIAMS D.M., TANAKA J. & NAGUMO T. 2006. ROUND F.E., HALLSTEINSEN H. & PAASCHE E. 1999. On a Morphological investigations of the frustule, perizonium and previously controversial ‘‘fragilarioid’’ diatom now placed in a initial valves of the freshwater diatom Achnanthes crenulata new genus Nanofrustulum. Diatom Research 14: 343–356. Grunow (Bacillariophyceae). Phycological Research 54: 173–182. SATO S., NAGUMO T. & TANAKA J. 2004. Auxospore formation and TROBAJO R., MANN D.G., CHEPURNOV V.A., CLAVERO E. & COX the morphology of the initial cell of the marine araphid diatom E.J. 2006. Auxosporulation and size reduction pattern in Nitzschia Gephyria media (Bacillariophyceae). Journal of Phycology 40: fonticola (Bacillariophyta). Journal of Phycology 42: 1353–1372. 684–691. UNDERWOOD G.J.C. & KROMKAMP J. 1999. Primary production by SATO S., MANN D.G., NAGUMO T., TANAKA J. & MEDLIN L.K. phytoplankton and microphytobenthos in estuaries. Advances in 2008a. Life cycle of Grammatophora marina (Bacillariophyta) Ecological Research 29: 93–153. with special reference to its auxospore fine structure. Phycologia VAN DE PEER, Y., CAERS A., DE RIJK P. & DE WACHTER R. 1998. 47: 12–27. Database on the structure of small ribosomal subunit RNA. SATO S., KURIYAMA K., TADANO T. & MEDLIN K.L. 2008b. Nucleic Acids Research 26: 179–182. Auxospore fine structure in a marine araphid diatom Tabularia VAN DE PEER Y., NEEFS J.-M., DE RIJK P. & DE WACHTER R. parva (Bacillariophyta). Diatom Research. In press. 1993. Reconstructing evolution from eukaryotic small-ribosom- SCHMID A.-M.M. 1994. Aspects of morphogenesis and function of al-subunit RNA sequences: calibration of the molecular clock. diatom cell walls with implications for taxonomy. Protoplasma Journal of Molecular Evolution 37: 221–232. 181: 43–60. VAN LANDINGHAM S.L. 1978. Catalogue of the fossil and recent genera SCHMID A.-M.M. & CRAWFORD R.M. 2001. Ellerbeckia arenaria and species of diatoms and their synonyms. Part VII. Rhoico- (Bacillariophyceae): formation of auxospores and initial cells. sphenia through Zygoceros. J. Cramer, Vaduz, Liechtenstein. European Journal of Phycology 36: 307–320. vON STOSCH H.A. 1962. U¨ ber das Perizonium der Diatomeen. SHULL V.L., VOGLER A.P., BAKER M.D., MADDISON D.R. & Vortra¨ge aus dem Gesamtgebiet der Botanik 1: 43–52. HAMMOND P.M. 2001. Sequence alignment of 18S ribosomal vON STOSCH H.A. 1982. On auxospore envelopes in diatoms. RNA and the basal relationships of adephagan beetles: evidence Bacillaria 5: 127–156. for monophyly of aquatic families and the placement of vON STOSCH H.A. & KOWALLIK K.V. 1969. Der von Geitler Trachypachidae. Systematic Biology 50: 945–969. aufgestellte Satz u¨ber die Notwendigkeit einer Mitose fu¨r jede SIMS P.A. & HOLMES R.W. 1983. Studies on the ‘‘kittonii’’ group of Schalenbildung von Diatomeen. Beobachtung u¨ber die Reich- Aulacodiscus species. Bacillaria 6: 267–292. weite und U¨ berlegungen zu seiner zellmechanischen Bedeutung. ¨ SIMS P.A., MANN D.G. & MEDLIN L.K. 2006. Evolution of the Osterreichische botanische Zeitschrift 116: 454–474. diatoms: insights from fossil, biological and molecular data. vON STOSCH H.A., THEIL G. & KOWALLIK K.V. 1973. Entwick- Phycologia 45: 361–402. lungsgeschichtliche Untersuchungen an zentrischen Diatomeen. Chaetoceros didymum SORHANNUS U. 2004. Diatom phylogenetics inferred based on V. Bau und Lebenszyklus von ,mit direct optimization nuclear-encoded SSU rRNA sequences. Beobachtungen u¨ber einige andere Arten der Gattung. Helgo- Cladistics 20: 487–497. la¨nder wissenschaftliche Meeresuntersuchungen 25: 384–445. WITKOWSKI A., LANGE-BERTALOT H. & METZELTIN D. 2000. SORHANNUS U. 2007. A nuclear-encoded small-subunit ribosomal Diatom flora of marine coasts I. Iconographia Diatomologica 7: RNA timescale for diatom evolution. Marine Micropaleontology 1–925. 65: 1–12. STAMATAKIS A., LUDWIG T. & MEIER H. 2005. RAxML-III: a fast program for maximum likelihood-based inference of large Received 12 January 2008; accepted 18 February 2008 phylogenetic trees. Bioinformatics 21: 456–463. Associate editor: Wiebe Kooistra